Statistical analysis inappropriateBMJ 1994; 308 doi: https://doi.org/10.1136/bmj.308.6924.339b (Published 29 January 1994) Cite this as: BMJ 1994;308:339
- J M Bland
- Department of Public Health Sciences, St George's Hospital Medical School, London SW17 0RE.
EDITOR, - S C Darby and colleagues' study of cancer and atmospheric nuclear weapons tests raises some interesting methodological questions.1 Exposed servicemen and civilians working with nuclear weapons were compared with a control group matched for age and occupation and with the general population by means of national mortality statistics. Two sided tests and 95% confidence intervals were used to compare the exposed group with the general population. One sided tests and 90% confidence intervals were used to compare the exposed group with the control group. The tests were done in the direction of the observed difference.
In a one sided test the alternative hypothesis is that there is a difference in a specified direction. The null hypothesis is then that there …